
# Essential libraries for all parts of the data science lifecycle
import pandas as pd
import numpy as np

# for plotting data and visualizations
import matplotlib.pyplot as plt
import seaborn as sns

# for explanatory analysis section and creating many models for machine learning
from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error
from sklearn.ensemble import RandomForestRegressor
from sklearn.model_selection import train_test_split
from sklearn.metrics import r2_score
from sklearn import metrics
from sklearn.model_selection import cross_val_score, KFold
from sklearn.linear_model import Lasso
from sklearn.svm import SVR
from sklearn.multioutput import MultiOutputRegressor


# reading CSV through pandas
df = pd.read_csv('powerlifting.csv', dtype=object)

# displaying Dataframe
df


# get rid of unnecessary columns
df = df[['Date', 'Age', 'Sex', 'Event', 'BirthYearClass', 'BodyweightKg', \
       'Squat1Kg', 'Squat2Kg', 'Squat3Kg', 'Best3SquatKg', \
       'Bench1Kg', 'Bench2Kg', 'Bench3Kg', 'Best3BenchKg', \
       'Deadlift1Kg', 'Deadlift2Kg', 'Deadlift3Kg', \
       'Best3DeadliftKg', 'TotalKg', 'Place', 'Dots', 'Wilks', 'Glossbrenner', \
       'Tested', 'Federation']]


# drop all rows with Nan values since we only want to analyze people who completed all their lifts
# and have important data like their weight and age/birthyearclass
df = df.dropna()

# restrict to just USAPL and IPF federations
df = df[(df['Federation'] == 'USAPL') | (df['Federation'] == 'IPF')]

# restrict to SBD and Tested meets only
df = df[(df['Event'] == 'SBD') & (df['Tested'] == 'Yes')]
df


for i,row in df.iterrows():
    row['Date'] = pd.to_datetime(row['Date'])
    try:
        row['Age'] = float(row['Age'])
    except:
        row['Age'] = np.nan
    # want to nake distinction between Male and Female 
    if row['Sex'] == 'Mx':
        row['Sex'] = np.nan
        
    lifts = ['Squat1Kg', 'Squat2Kg', 'Squat3Kg', 'Best3SquatKg', \
       'Bench1Kg', 'Bench2Kg', 'Bench3Kg', 'Best3BenchKg', \
       'Deadlift1Kg', 'Deadlift2Kg', 'Deadlift3Kg', \
       'Best3DeadliftKg', 'TotalKg']
      
    # cast all lift values as floats and get rid of invalid data
    for lift in lifts:
        try:
            row[lift] = float(row[lift])
            if row[lift] <= 0:
                 row[lift] = np.nan
        except:
             row[lift] = np.nan
    # cast all places as integers and get rid of invalid data
    try:
        row['Place'] = int(row['Place'])
        if row['Place'] > 0:
            row['Place'] = int(row['Place'])
        else:
            row['Place'] = np.nan
    except:
        row['Place'] = np.nan
    
    # cast all scores as floats and get rid of invalid data         
    for score in ['Dots', 'Wilks', 'Glossbrenner']:
        try:
            row[score] = float(row[score])  
            if row[score] > 0:
                 row[score] = float(row[score])
            else:
                row[score] = np.nan
        except:
            row[score] = np.nan
cleanedData = df.dropna()
# We fixed the event to be SBD and the Federations to only be USAPL and IPF which are tested federations
cleanedData = cleanedData.drop(['Event', 'Tested', 'Federation'], axis=1)
cleanedData


#cleanedData = cleanedData.astype({'Date': 'datetime64[ns]', 'Age': 'float64', 'BodyweightKg': 'float64', \
#                                 'Squat1Kg': 'float64', 'Squat2Kg': 'float64', 'Squat3Kg': 'float64', 'Best3SquatKg': 'float64',\
#                                 'Bench1Kg': 'float64', 'Bench2Kg': 'float64', 'Bench3Kg': 'float64', 'Best3BenchKg': 'float64',\
#                                 'Deadlift1Kg': 'float64', 'Deadlift2Kg': 'float64', 'Deadlift3Kg': 'float64', 'Best3DeadliftKg': 'float64',\
#                                 'TotalKg': 'float64', 'Place': 'int', 'Dots': 'float64', 'Wilks': 'float64',  'Glossbrenner': 'float64'})

#cleanedData.dtypes

# save cleaned data so that I can do analysis without having to run all above cells on a kernal restart
#cleanedData.to_csv('cleaned_data.csv', index=False)


# reading CSV through pandas
df = pd.read_csv('cleaned_data.csv')

# displaying head of Dataframe
df.head()


import matplotlib.pyplot as plt

# Create Weight Classes
bins = [0, 52, 57, 63, 72, 84, 93, 105, 120, 1200]
labels = ['47kg', '52kg', '57kg', '63kg', '72kg', '84kg', '93kg', '105kg', '120kg+']
df['WeightClass'] = pd.cut(df['BodyweightKg'], bins=bins, labels=labels)

# Filter observations by gender
male_df = df[df['Sex'] == 'M']
female_df = df[df['Sex'] == 'F']

# Get count of observations by weight class for each gender
male_weightclass_counts = male_df['WeightClass'].value_counts()
female_weightclass_counts = female_df['WeightClass'].value_counts()

# Sort weight class categories
sorted_weightclasses = sorted(labels, key=lambda x: labels.index(x))

# Plot count of observations by weight class for males
plt.figure(figsize=(10, 4))  # Adjust the figure size as needed
plt.subplot(1, 2, 1)
male_weightclass_counts.loc[sorted_weightclasses].plot(kind='bar')
plt.title('Count of Observations by Weight Class (Male)')
plt.xlabel('Weight Class')
plt.ylabel('Count')

# Plot count of observations by weight class for females
plt.subplot(1, 2, 2)
female_weightclass_counts.loc[sorted_weightclasses].plot(kind='bar')
plt.title('Count of Observations by Weight Class (Female)')
plt.xlabel('Weight Class')
plt.ylabel('Count')

plt.tight_layout()  # Adjust spacing between subplots if needed
plt.show()


# create ordered categories for birth year class
ordered_birthyear = pd.Categorical(df['BirthYearClass'], 
                                    categories=['14-18', '19-23', '24-39', '40-49', '50-59', '60-69', '70-999'],
                                    ordered=True)

# create a copy of the dataframe with ordered birth year class
df_ordered = df.copy()
df_ordered['BirthYearClass'] = ordered_birthyear
df_ordered = df_ordered.sort_values(by='BirthYearClass')

# create separate plots for males and females
male_df = df_ordered[df_ordered['Sex'] == 'M']
female_df = df_ordered[df_ordered['Sex'] == 'F']

fig, (ax1, ax2) = plt.subplots(ncols=2, figsize=(16, 6))

# plot for males
sns.boxplot(x='BirthYearClass', y='TotalKg', data=male_df, ax=ax1)
ax1.set_title('Males')

# plot for females
sns.boxplot(x='BirthYearClass', y='TotalKg', data=female_df, ax=ax2)
ax2.set_title('Females')

plt.tight_layout()
plt.show()


# for future use when wanting to compare males and females
male_lifters = df[df['Sex'] == 'M']
female_lifters = df[df['Sex'] == 'F']


# Create subplots
fig, axes = plt.subplots(ncols=2, figsize=(15, 6))

# Plot male data
axes[0].scatter(male_lifters['BodyweightKg'], male_lifters['Dots'])
axes[0].set_xlabel('Bodyweight (kg)')
axes[0].set_ylabel('Dots Score')
axes[0].set_title('Male Lifters')

# Plot female data
axes[1].scatter(female_lifters['BodyweightKg'], female_lifters['Dots'])
axes[1].set_xlabel('Bodyweight (kg)')
axes[1].set_ylabel('Dots Score')
axes[1].set_title('Female Lifters')

y_ticks = [0, 100, 200, 300, 400, 500, 600, 700]
axes[0].set_yticks(y_ticks)
axes[1].set_yticks(y_ticks)

plt.tight_layout()
plt.show()


fig, axs = plt.subplots(2, 3, figsize=(20, 10))
fig.subplots_adjust(hspace=0.5)

for i, sex in enumerate(['M', 'F']):
    df_sex = df[df['Sex'] == sex]
    for j, lift in enumerate(['Squat', 'Bench', 'Deadlift']):
        col_lift = lift + '2Kg'
        col_total = 'TotalKg'
        axs[i, j].scatter(df_sex[col_lift], df_sex[col_total], alpha=0.5)
        axs[i, j].set_title(sex + ' ' + lift)
        axs[i, j].set_xlabel('Second Attempt ' + lift)
        axs[i, j].set_ylabel('Total')
plt.show()


# group by sex and birth year class
grouped = df.groupby(['Sex', 'BirthYearClass'])

# iterate over groups and plot correlations
for i, (group, data) in enumerate(grouped):
    sex, birthyearclass = group
    
    # get the middle of each birth year class
    birthyearclass_mid = [int((int(x.split("-")[0]) + int(x.split("-")[1]))/2) for x in data['BirthYearClass']]
    
    # plot correlations
    fig, axs = plt.subplots(nrows=1, ncols=3, figsize=(15,5))
    
    axs[0].scatter(data['Bench1Kg'], data['TotalKg'], alpha=0.5)
    axs[0].set_xlabel('Bench 1st Attempt (kg)')
    axs[0].set_ylabel('Total Weight (kg)')
    
    axs[1].scatter(data['Squat1Kg'], data['TotalKg'], alpha=0.5)
    axs[1].set_xlabel('Squat 1st Attempt (kg)')
    axs[1].set_ylabel('Total Weight (kg)')
    
    axs[2].scatter(data['Deadlift1Kg'], data['TotalKg'], alpha=0.5)
    axs[2].set_xlabel('Deadlift 1st Attempt (kg)')
    axs[2].set_ylabel('Total Weight (kg)')
    
    
    fig.suptitle(f'{sex} {birthyearclass} - Correlation of First Attempts and Total Weight')
    
    plt.tight_layout()
    plt.show()


# Calculate percentage difference for attempts 1 to 2 and 2 to 3
df['Bench12PercentDiff'] = ((df['Bench2Kg'] - df['Bench1Kg']) / df['Bench1Kg']) * 100
df['Bench23PercentDiff'] = ((df['Bench3Kg'] - df['Bench2Kg']) / df['Bench2Kg']) * 100
df['Squat12PercentDiff'] = ((df['Squat2Kg'] - df['Squat1Kg']) / df['Squat1Kg']) * 100
df['Squat23PercentDiff'] = ((df['Squat3Kg'] - df['Squat2Kg']) / df['Squat2Kg']) * 100
df['Deadlift12PercentDiff'] = ((df['Deadlift2Kg'] - df['Deadlift1Kg']) / df['Deadlift1Kg']) * 100
df['Deadlift23PercentDiff'] = ((df['Deadlift3Kg'] - df['Deadlift2Kg']) / df['Deadlift2Kg']) * 100

# Group the data by lift and calculate mean percentage difference for each lift
lift_groups = df.groupby('WeightClass')[['Bench12PercentDiff', 'Bench23PercentDiff', 'Squat12PercentDiff', 'Squat23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']]
mean_percent_diff = lift_groups.mean()
mean_percent_diff


# Create bar plots of mean percentage difference for each lift
lifts = ['Bench', 'Squat', 'Deadlift']
lift_diffs = ['12PercentDiff', '23PercentDiff']
fig, axs = plt.subplots(len(lifts), len(lift_diffs), figsize=(12, 8))
y_ticks = [0,1,2,3,4,5,6,7,8,9,10]
for i, lift in enumerate(lifts):
    for j, diff in enumerate(lift_diffs):
        col_name = f"{lift}{diff}"
        axs[i, j].bar(mean_percent_diff.index, mean_percent_diff[col_name])
        axs[i, j].set_title(f"{lift} {diff} mean percentage difference")
        axs[i, j].set_xlabel('Weight class')
        axs[i, j].set_ylabel('Percentage difference')
        axs[i, j].set_yticks(y_ticks)
fig.tight_layout()
plt.show()


lifts = ['Squat', 'Bench', 'Deadlift']
lift_diff_cols = ['Squat12PercentDiff', 'Squat23PercentDiff', 'Bench12PercentDiff', 'Bench23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']
titles = ["Bench Attempt 1 to 2 % Diff vs Total (by Birth Class)",
          "Bench Attempt 2 to 3 % Diff vs Total (by Birth Class)",
          "Squat Attempt 1 to 2 % Diff vs Total (by Birth Class)",
          "Squat Attempt 2 to 3 % Diff vs Toal (by Birth Class)",
          "Deadlift Attempt 1 to 2 % Diff vs Total (by Birth Class)",
          "Deadlift Attempt 2 to 3 % Diff vs Total (by Birth Class)"]
x_titles = ["Bench Attempt 1 to 2 Percent Difference",
           "Bench Attempt 2 to 3 Percent Difference",
           "Squat Attempt 1 to 2 % Diff",
          "Squat Attempt 2 to 3 Percent Difference",
          "Deadlift Attempt 1 to 2 Percent Difference",
          "Deadlift Attempt 2 to 3 Percent Difference"]
total_col = 'TotalKg'
byc_col = 'BirthYearClass'

fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(15,12))
fig.subplots_adjust(hspace=2)

for i, lift in enumerate(lifts):
    lift_diff_cols_for_lift = [col for col in lift_diff_cols if lift in col]
    for j, diff_col in enumerate(lift_diff_cols_for_lift):
        # Create a scatter plot of Total vs percent difference, split by birth year class
        ax = axs[i,j]
        for byc in df[byc_col].unique():
            mask = df[byc_col] == byc
            x = df.loc[mask, diff_col]
            y = df.loc[mask, total_col]
            ax.scatter(x, y, label=byc)
        ax.set_xlabel(x_titles[i*2+j])
        ax.set_ylabel(total_col)
        ax.set_title(titles[i*2+j])
        ax.set_xticks(range(0, 226, 25))
        ax.legend(title=byc_col, loc='upper right')
        
plt.tight_layout()
plt.show()


fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(15,12))
fig.subplots_adjust(hspace=2)

for i, lift in enumerate(lifts):
    lift_diff_cols_for_lift = [col for col in lift_diff_cols if lift in col]
    for j, diff_col in enumerate(lift_diff_cols_for_lift):
        # Create a scatter plot of Total vs percent difference, split by birth year class
        ax = axs[i,j]
        for byc in df[byc_col].unique():
            mask = df[byc_col] == byc
            x = df.loc[mask, diff_col]
            y = df.loc[mask, total_col]
            ax.scatter(x, y, label=byc)
        ax.set_xlabel(x_titles[i*2+j])
        ax.set_ylabel(total_col)
        ax.set_title(titles[i*2+j])
        ax.set_xlim(0, 30)
        ax.set_xticks(range(0, 31, 5))
        ax.legend(title=byc_col, loc='upper right')
        
plt.tight_layout()
plt.show()


#new_df = df.drop(['Date','Glossbrenner', 'Dots', 'Place'], axis=1)
#new_df.to_csv('ML_data.csv', index=False)


# reading CSV through pandas
df = pd.read_csv('ML_data.csv')

# displaying head of Dataframe
df.head()


# Split the data into features (X) and target (y)
X = df[['TotalKg', 'Squat1Kg', 'Bench1Kg', 'Deadlift1Kg']]
y = df[['Bench12PercentDiff', 'Bench23PercentDiff', 'Squat12PercentDiff',
       'Squat23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']].values

# Split the data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.2, random_state=42)

# Initialize a linear regression model
lr_model = LinearRegression()

# Fit the model to the training data
lr_model.fit(X_train, y_train)

# Make predictions on the testing data
y_pred = lr_model.predict(X_test)

# Print the model's score (R^2)
print(f"R^2 Score: {lr_model.score(X_test, y_test)}")

# Calculate and print other evaluation metrics
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
print('Explained Variance Score:', metrics.explained_variance_score(y_test, y_pred))

R^2 Score: 0.4645126873207009
Mean Absolute Error: 1.6326704761320168
Mean Squared Error: 6.713648768307856
Root Mean Squared Error: 2.591070969369202
Explained Variance Score: 0.4645853619067515


# Create a scatter plot of actual vs predicted values
fig = plt.figure(figsize=(8, 6))
plt.scatter(y_test, y_pred)

# Calculate the coefficients of the regression line
coefficients = np.polyfit(y_test.flatten(), y_pred.flatten(), 1)

# Calculate the regression line values
regression_x = np.array([np.min(y_test), np.max(y_test)])
regression_y = np.polyval(coefficients, regression_x)

# Add the regression line to the scatter plot
plt.plot(regression_x, regression_y, '--', color='red', label=f'Regression Line: y = {coefficients[0]:.2f}x + {coefficients[1]:.2f}')

# Add labels and legend
plt.xlabel('Actual')
plt.ylabel('Predicted')
plt.title('Actual vs. Predicted Values')
plt.legend()
plt.xlim(0,35)

# Show the plot
plt.show()


# Split the data into features (X) and target (y)
X = df[['TotalKg', 'Squat1Kg', 'Bench1Kg', 'Deadlift1Kg']]
y = df[['Bench12PercentDiff', 'Bench23PercentDiff', 'Squat12PercentDiff',
       'Squat23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']].values

# Split the data into training and testing sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=.2, random_state=42)

# Initialize a random forest model with 100 trees
rf_model = RandomForestRegressor(n_estimators=100, random_state=42)

# Fit the model to the training data
rf_model.fit(X_train, y_train)

# Make predictions on the testing data
y_pred = rf_model.predict(X_test)

# Print the model's score (R^2)
print(f"R^2 Score: {rf_model.score(X_test, y_test)}")

# Calculate and print other evaluation metrics
print('Mean Absolute Error:', metrics.mean_absolute_error(y_test, y_pred))
print('Mean Squared Error:', metrics.mean_squared_error(y_test, y_pred))
print('Root Mean Squared Error:', np.sqrt(metrics.mean_squared_error(y_test, y_pred)))
print('Explained Variance Score:', metrics.explained_variance_score(y_test, y_pred))

R^2 Score: 0.522873224046696
Mean Absolute Error: 1.506278644827509
Mean Squared Error: 5.903331255768655
Root Mean Squared Error: 2.429677191679721
Explained Variance Score: 0.5229470055435405


sns.set(style='white', font_scale=1.5)

# Create subplots with shared x-axis
fig, axs = plt.subplots(nrows=2, sharex=True, figsize=(10,8))

# Create scatterplot with regression line for actual vs. predicted values
sns.regplot(x=y_test.flatten(), y=y_pred.flatten(), ax=axs[0], line_kws={"color":"red"})
axs[0].set_title('Actual vs. Predicted')
axs[0].set_xlabel('Actual')
axs[0].set_ylabel('Predicted')
axs[0].set_ylim([0,70])

# Create density plot for actual vs. predicted values
sns.kdeplot(y_test.flatten(), label='Actual', fill=True, ax=axs[1])
sns.kdeplot(y_pred.flatten(), label='Predicted', fill=True, ax=axs[1])
axs[1].set_title('Density Plot')
axs[1].set_xlabel('Percent Difference')
axs[1].set_ylabel('Density')
axs[1].set_ylim([0,0.4])

# Set x-axis limits for both subplots
for ax in axs:
    ax.set_xlim([0, 60])
# Show the plot
plt.subplots_adjust(hspace=0.5)
plt.show()


# Set up the data
X = df[['TotalKg', 'Age', 'BodyweightKg', 'Squat1Kg', 'Bench1Kg', 'Deadlift1Kg']].values
y = df[['Bench12PercentDiff', 'Bench23PercentDiff', 'Squat12PercentDiff',       'Squat23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']].values

# Find the best k value using k-fold cross-validation
best_score = -np.inf
r2_scores = []
k_values = [2,3,4,5,6,7,8,9,10,11]
for k in k_values:
    kf = KFold(n_splits=k, shuffle=True, random_state=42)
    scores = []
    for train_idx, test_idx in kf.split(X):
        X_train, X_test = X[train_idx], X[test_idx]
        y_train, y_test = y[train_idx], y[test_idx]
        rf_model = RandomForestRegressor(n_estimators=100, random_state=42)
        rf_model.fit(X_train, y_train)
        score = rf_model.score(X_test, y_test)
        scores.append(score)
    avg_score = np.mean(scores)
    r2_scores.append(avg_score)
    print(f"Average R^2 score for k={k}: {avg_score}")

Average R^2 score for k=2: 0.4765926787770881
Average R^2 score for k=3: 0.5001842960634079
Average R^2 score for k=4: 0.50599654592396
Average R^2 score for k=5: 0.5214194230165615
Average R^2 score for k=6: 0.5243440259081164
Average R^2 score for k=7: 0.5245099152857321
Average R^2 score for k=8: 0.5273002718337676
Average R^2 score for k=9: 0.5340681973128207
Average R^2 score for k=10: 0.5369092554835346
Average R^2 score for k=11: 0.534841326223508


plt.plot(k_values, r2_scores)
plt.xlabel('k value')
plt.ylabel('R^2 Score')
plt.show()


# Set up the data
X = df[['TotalKg', 'Age', 'BodyweightKg', 'Squat1Kg', 'Bench1Kg', 'Deadlift1Kg']].values
y = df[['Bench12PercentDiff', 'Bench23PercentDiff', 'Squat12PercentDiff', 'Squat23PercentDiff', 'Deadlift12PercentDiff', 'Deadlift23PercentDiff']].values

# Train the random forest model using KFolds with k=10 and evaluate its performance
kf = KFold(n_splits=10, shuffle=True, random_state=42)
scores = []
mae_scores = []
mse_scores = []
rmse_scores = []
ev_scores = []
tests = []
predictions = []
for train_idx, test_idx in kf.split(X):
    X_train, X_test = X[train_idx], X[test_idx]
    y_train, y_test = y[train_idx], y[test_idx]
    tests.append(y_test)
    rf_model = RandomForestRegressor(n_estimators=100, random_state=42)
    rf_model.fit(X_train, y_train)
    y_pred = rf_model.predict(X_test)
    score = rf_model.score(X_test, y_test)
    mae = metrics.mean_absolute_error(y_test, y_pred)
    mse = metrics.mean_squared_error(y_test, y_pred)
    rmse = np.sqrt(mse)
    ev = metrics.explained_variance_score(y_test, y_pred)
    scores.append(score)
    mae_scores.append(mae)
    mse_scores.append(mse)
    rmse_scores.append(rmse)
    ev_scores.append(ev)
    predictions.append(y_pred)
avg_score = np.mean(scores)
avg_mae = np.mean(mae_scores)
avg_mse = np.mean(mse_scores)
avg_rmse = np.mean(rmse_scores)
avg_ev = np.mean(ev_scores)
print(f"Average R^2 score: {avg_score}")
print(f"Mean Absolute Error: {avg_mae}")
print(f"Mean Squared Error: {avg_mse}")
print(f"Root Mean Squared Error: {avg_rmse}")
print(f"Explained Variance Score: {avg_ev}")

Average R^2 score: 0.5369092554835346
Mean Absolute Error: 1.4530920899570874
Mean Squared Error: 6.138609798944555
Root Mean Squared Error: 2.472332262131271
Explained Variance Score: 0.5371134551961355


# Create a scatter plot of actual vs predicted values
# Convert generator object to list
min_len = min([len(pred) for pred in predictions])
for i in range(len(predictions)):
    predictions[i] = predictions[i][:min_len]
    tests[i] = tests[i][:min_len]

# Compute the mean of the predictions and the actual values
y_pred = np.mean(predictions, axis=0)
y_test = np.mean(tests, axis=0)

sns.set(style='white', font_scale=1.5)

# Create subplots with shared x-axis
fig, axs = plt.subplots(nrows=2, sharex=True, figsize=(10,8))

# Create scatterplot with regression line for actual vs. predicted values
sns.regplot(x=y_test.flatten(), y=y_pred.flatten(), ax=axs[0], line_kws={"color":"red"})
axs[0].set_title('Actual vs. Predicted')
axs[0].set_xlabel('Actual')
axs[0].set_ylabel('Predicted')
axs[0].set_ylim([0,40])

# Create density plot for actual vs. predicted values
sns.kdeplot(y_test.flatten(), label='Actual', fill=True, ax=axs[1])
sns.kdeplot(y_pred.flatten(), label='Predicted', fill=True, ax=axs[1])
axs[1].set_title('Density Plot')
axs[1].set_xlabel('Percent Difference')
axs[1].set_ylabel('Density')
axs[1].set_ylim([0,0.4])

# Set x-axis limits for both subplots
for ax in axs:
    ax.set_xlim([0, 30])
# Show the plot
plt.subplots_adjust(hspace=0.5)
plt.show()

	Name	Sex	Event	Equipment	Age	AgeClass	BirthYearClass	Division	BodyweightKg	WeightClassKg	...	Tested	Country	State	Federation	ParentFederation	Date	MeetCountry	MeetState	MeetTown	MeetName
0	Alona Vladi	F	SBD	Raw	33	24-34	24-39	O	58.3	60	...	Yes	Russia	NaN	GFP	NaN	2019-05-11	Russia	NaN	Bryansk	Open Tournament
1	Galina Solovyanova	F	SBD	Raw	43	40-44	40-49	M1	73.1	75	...	Yes	Russia	NaN	GFP	NaN	2019-05-11	Russia	NaN	Bryansk	Open Tournament
2	Daniil Voronin	M	SBD	Raw	15.5	16-17	14-18	T	67.4	75	...	Yes	Russia	NaN	GFP	NaN	2019-05-11	Russia	NaN	Bryansk	Open Tournament
3	Aleksey Krasov	M	SBD	Raw	35	35-39	24-39	O	66.65	75	...	Yes	Russia	NaN	GFP	NaN	2019-05-11	Russia	NaN	Bryansk	Open Tournament
4	Margarita Pleschenkova	M	SBD	Raw	26.5	24-34	24-39	O	72.45	75	...	Yes	Russia	NaN	GFP	NaN	2019-05-11	Russia	NaN	Bryansk	Open Tournament
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2887194	Inna Orobets	F	SBD	Single-ply	45	45-49	40-49	Open	109.3	84+	...	Yes	Ukraine	NaN	IPF	IPF	2016-11-14	USA	NaN	Orlando	World Open Championships
2887195	Ágnes Szabó	F	SBD	Single-ply	27.5	24-34	24-39	Open	113.22	84+	...	Yes	Hungary	NaN	IPF	IPF	2016-11-14	USA	NaN	Orlando	World Open Championships
2887196	Brenda van der Meulen	F	SBD	Single-ply	37.5	35-39	24-39	Open	110.79	84+	...	Yes	Netherlands	NaN	IPF	IPF	2016-11-14	USA	NaN	Orlando	World Open Championships
2887197	Kelly Phasey	F	SBD	Single-ply	39.5	35-39	40-49	Open	124.53	84+	...	Yes	England	NaN	IPF	IPF	2016-11-14	USA	NaN	Orlando	World Open Championships
2887198	Hildeborg Juvet Hugdal	F	SBD	Single-ply	32.5	24-34	24-39	Open	123	84+	...	Yes	Norway	NaN	IPF	IPF	2016-11-14	USA	NaN	Orlando	World Open Championships

	Date	Age	Sex	Event	BirthYearClass	BodyweightKg	Squat1Kg	Squat2Kg	Squat3Kg	Best3SquatKg	...	Deadlift2Kg	Deadlift3Kg	Best3DeadliftKg	TotalKg	Place	Dots	Wilks	Glossbrenner	Tested	Federation
307302	2018-11-10	48	F	SBD	40-49	51.35	102.5	110	112.5	112.5	...	122.5	127.5	127.5	292.5	1	359.67	368.19	326.53	Yes	USAPL
307304	2018-11-10	47	F	SBD	40-49	71.6	100	112.5	120	120	...	142.5	150	150	325	1	324.56	318.39	280.50	Yes	USAPL
307305	2018-11-10	43.5	F	SBD	40-49	114.65	70	80	92.5	92.5	...	127.5	140	140	315	1	256.19	254.00	215.78	Yes	USAPL
307306	2018-11-10	38.5	F	SBD	24-39	51.1	92.5	95	97.5	97.5	...	117.5	125	125	275	1	339.31	347.46	308.21	Yes	USAPL
307307	2018-11-10	26.5	F	SBD	24-39	57	92.5	97.5	102.5	102.5	...	112.5	115	115	277.5	1	317.93	322.01	284.73	Yes	USAPL
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2887193	2016-11-14	32	F	SBD	24-39	95	250	255	260	260	...	222.5	227.5	227.5	647.5	2	564.05	548.04	473.70	Yes	IPF
2887194	2016-11-14	45	F	SBD	40-49	109.3	235	-242.5	242.5	242.5	...	210	215	215	645	3	532.83	525.14	447.12	Yes	IPF
2887195	2016-11-14	27.5	F	SBD	24-39	113.22	-237.5	237.5	-247.5	237.5	...	200	-210	200	632.5	4	516.46	511.26	434.61	Yes	IPF
2887196	2016-11-14	37.5	F	SBD	24-39	110.79	210	222.5	227.5	227.5	...	190	-195	190	617.5	5	507.79	501.32	426.60	Yes	IPF
2887197	2016-11-14	39.5	F	SBD	40-49	124.53	210	220	-227.5	220	...	-155	-155	145	515	6	409.13	409.13	345.98	Yes	IPF

	Date	Age	Sex	BirthYearClass	BodyweightKg	Squat1Kg	Squat2Kg	Squat3Kg	Best3SquatKg	Bench1Kg	...	Best3BenchKg	Deadlift1Kg	Deadlift2Kg	Deadlift3Kg	Best3DeadliftKg	TotalKg	Place	Dots	Wilks	Glossbrenner
307305	2018-11-10 00:00:00	43.5	F	40-49	114.65	70.0	80.0	92.5	92.5	65.0	...	82.5	110.0	127.5	140.0	140.0	315.0	1	256.19	254.0	215.78
307309	2018-11-10 00:00:00	24.5	F	24-39	58.95	102.5	110.0	115.0	115.0	57.5	...	62.5	127.5	135.0	142.5	142.5	320.0	2	358.73	361.69	319.65
307314	2018-11-10 00:00:00	24.5	F	24-39	71	110.0	117.5	122.5	122.5	67.5	...	75.0	137.5	147.5	155.0	155.0	352.5	3	353.66	347.3	306.03
307318	2018-11-10 00:00:00	30.5	F	24-39	97.15	150.0	165.0	172.5	172.5	65.0	...	77.5	165.0	175.0	187.5	187.5	437.5	1	377.64	367.52	316.84
307319	2018-11-10 00:00:00	43.5	F	40-49	114.65	70.0	80.0	92.5	92.5	65.0	...	82.5	110.0	127.5	140.0	140.0	315.0	2	256.19	254.0	215.78
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
2886975	2022-08-27 00:00:00	18.5	F	19-23	69.9	155.0	165.0	172.5	172.5	75.0	...	82.5	140.0	147.5	150.0	150.0	405.0	3	409.89	403.31	355.52
2886978	2022-08-27 00:00:00	20.5	F	19-23	82.3	150.0	165.0	175.0	175.0	100.0	...	112.5	130.0	140.0	150.0	150.0	437.5	3	406.53	394.24	345.14
2886998	2016-11-14 00:00:00	37.5	M	24-39	65.8	250.0	260.0	262.5	262.5	190.0	...	200.0	275.0	287.5	300.0	300.0	762.5	3	598.79	600.21	583.31
2887072	2016-11-14 00:00:00	31.0	M	24-39	118.25	400.0	415.0	422.5	422.5	310.0	...	327.5	347.5	360.0	375.0	375.0	1125.0	1	649.29	649.05	621.82
2887115	2016-11-14 00:00:00	39.5	F	40-49	51.59	155.0	165.0	175.0	175.0	90.0	...	100.0	150.0	160.0	165.0	165.0	440.0	6	539.28	551.88	489.34

	Date	Age	Sex	BirthYearClass	BodyweightKg	Squat1Kg	Squat2Kg	Squat3Kg	Best3SquatKg	Bench1Kg	...	Best3BenchKg	Deadlift1Kg	Deadlift2Kg	Deadlift3Kg	Best3DeadliftKg	TotalKg	Place	Dots	Wilks	Glossbrenner
0	2018-11-10	43.5	F	40-49	114.65	70.0	80.0	92.5	92.5	65.0	...	82.5	110.0	127.5	140.0	140.0	315.0	1	256.19	254.00	215.78
1	2018-11-10	24.5	F	24-39	58.95	102.5	110.0	115.0	115.0	57.5	...	62.5	127.5	135.0	142.5	142.5	320.0	2	358.73	361.69	319.65
2	2018-11-10	24.5	F	24-39	71.00	110.0	117.5	122.5	122.5	67.5	...	75.0	137.5	147.5	155.0	155.0	352.5	3	353.66	347.30	306.03
3	2018-11-10	30.5	F	24-39	97.15	150.0	165.0	172.5	172.5	65.0	...	77.5	165.0	175.0	187.5	187.5	437.5	1	377.64	367.52	316.84
4	2018-11-10	43.5	F	40-49	114.65	70.0	80.0	92.5	92.5	65.0	...	82.5	110.0	127.5	140.0	140.0	315.0	2	256.19	254.00	215.78

	Bench12PercentDiff	Bench23PercentDiff	Squat12PercentDiff	Squat23PercentDiff	Deadlift12PercentDiff	Deadlift23PercentDiff
WeightClass
47kg	9.217058	6.686350	8.525675	6.063309	8.412385	6.047253
52kg	8.637485	6.005720	7.613078	5.711908	8.041704	5.826675
57kg	8.309339	5.893291	7.440370	5.652076	7.760215	5.851623
63kg	8.021261	5.614443	7.222643	5.475819	7.480750	5.596313
72kg	7.252966	5.068700	6.784184	5.124431	7.141596	5.246906
84kg	6.957221	4.861000	6.741862	5.020052	7.114098	5.193180
93kg	6.789667	4.866404	6.595476	5.010194	7.029945	5.096721
105kg	7.081237	4.910506	6.737796	5.107322	7.068021	5.224477
120kg+	6.875221	5.079433	6.754716	5.135136	7.075729	5.154207

A Data Science Dive Into Powerlifting Meets ¶

Exploring the Data Science Lifecycle ¶

Elias Prieto ¶

Introduction ¶

Why is this important?¶

Data Collection ¶

Part 1: Retrieving the Data¶

Part 2: Tidying the Data¶

Explantory Analysis and Data Visualization ¶

Modeling the Data ¶

Interpretation: Insight & Policy Decision ¶

Resources and Other Cool Projects ¶